MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Vol. 207: 89–96, 2000
Published November 22
Migration of a marine oligochaete: induction of
dispersal and microhabitat choice
Per G. Nilsson*, Jeffrey S. Levinton, Josepha P. Kurdziel
Department of Ecology and Evolution, State University of New York at Stony Brook, Stony Brook, New York 11794, USA
ABSTRACT: We present evidence for migration in the asexual phase of the life cycle of Paranais
litoralis (Müller), a marine oligochaete that reproduces asexually by fission. Depletion of resources
triggers a swimming response by some worms (‘migrators’). Migrating worms are longer, thinner and
have more segments than non-migrating worms, indicating that migrating worms postpone reproduction. Experiments show that migrators swim more actively than non-migrators, and that this
causes them to stay suspended in the water column longer than non-migrators. Worms avoid sediment where resources are exhausted, both when burrowing and when returning to the bottom after
swimming. Based on laboratory cultures and field observations, we hypothesize that local depletion
of resources is a common phenomenon for P. litoralis, and that migration is important for the persistence of P. litoralis populations.
KEY WORDS: Migration · Swimming · Colonization · Asexual reproduction · Spatial population
model
Resale or republication not permitted without written consent of the publisher
INTRODUCTION
The spatial dynamics of an organism may influence
the biology of an individual and of the population in
many ways (Dias 1996, Harrison & Hastings 1996,
Palmer et al. 1996, Sugg et al. 1996). Several authors
have noted that species living in ephemeral habitats
tend to migrate more than species living in stable habitats (e.g., Dingle 1996, Palmer et al. 1996). When migrators reach a new patch, they may remain there to start a
new population. Although populations may have a positive growth rate once established, demographic stochasticity during the establishment phase following
immigration may lead to frequent extinction (Nunney &
Campbell 1993). This may explain the absence of species from an area, despite suitable habitat and adequate dispersal capability. Two ways of decreasing the
risk of stochastic demographic extinction are to grow
fast (short generation time and high fecundity) and to
escape the need for a sexual partner. These are charac*Present address: Tjärnö Marine Biological Laboratory, 45296
Strömstad, Sweden. E-mail: [email protected]
© Inter-Research 2000
teristics of most asexual organisms, and (facultatively)
asexual organisms are common in ephemeral habitats
(Hughes 1989). However, asexual reproduction may
also have the potential disadvantage of not producing a
resting and/or migratory stage (Hughes 1989). Many
clonal organisms reproduce asexually while they can
expand locally, but when all locally available space or
resource is used they produce sexual offspring that constitute the migratory stage. It is, however, not necessary
for the migratory stage to be sexual: a parthenogenetically produced offspring in the form of a migratory
larva would still give the advantage of asexual reproduction. The fact that the migratory stage is so frequently sexual suggests the importance of genetic variation in the offspring, presumably because the larvae
may encounter conditions different from those experienced by the mother. If, however, the adverse conditions triggering the response to migrate are likely to be
spatially restricted, then the conditions at the place
where the migrant settles may still be similar to the conditions experienced by the mother, and an asexual dispersal phase may be more advantageous than a sexually produced offspring.
90
Mar Ecol Prog Ser 207: 89–96, 2000
Oligochaetes frequently occur in ephemeral and disturbed habitats, which suggests the presence of an
efficient dispersal mechanism. However, many species
of oligochaetes reproduce asexually by fission, especially in the families Naididae and Aeolosomatidae
(Lasserre 1975). When they reproduce sexually, it is
by insemination (possibly reciprocal) and by the formation of cocoons (Lasserre 1975, Giere & Pfannkuche
1982). The offspring emerge from the cocoon as a
smaller but morphologically similar version of the parents, i.e., oligochaetes lack a planktonic larval migratory stage. For example, the naidid oligochaete Paranais litoralis (Müller) is common worldwide in estuarine
habitats, especially intertidal mud flats (Giere &
Pfannkuche 1982). They typically undergo a pronounced yearly cycle, with low abundance during
winter, increasing population densities during spring,
reaching maximum population densities in early summer, after which the populations crash and occur only
at low densities until the following spring (Giere &
Pfannkuche 1982, Cheng et al. 1993, Cheng & Chang
1999). Sometimes a second but smaller population
increase is found in autumn. The population distribution is patchy during peak numbers (Nilsson et al.
1997). Laboratory cultures have shown that there are
differences in nutritional quality of the sediment both
among seasons and among places within seasons
(Cheng et al. 1993, Nilsson et al. 1997), consistent with
the patchy field occurrence of the oligochaetes. The
typical asexual offspring produced by P. litoralis is an
almost identical copy of the mother: in laboratory cultures, the offspring matures within about 4 d to produce its first offspring (Martinez 1993, Nilsson pers.
obs.). Since no recombination is involved in this process, it may also be seen as growth, although we use
the term asexual reproduction here to emphasize that
a new detached ‘unit’ is produced (Hughes 1989). We
have never observed sexual reproduction in P. litoralis
in our laboratory cultures, or sexually mature worms in
field samples. P. litoralis is also a typical oligochaete in
that it lacks a larval dispersal stage. However, we have
frequently observed adult individuals of P. litoralis
swimming in our high-density laboratory cultures. This
suggested to us that swimming might be a response by
the worms to adverse conditions, which may have
important effects on the population biology of P. litoralis if it occurs in nature. Many investigators have
reported normally infaunal meiofauna-sized animals in
the water column (Giere 1993). Several passive or
active triggering mechanisms for water-column dispersal have been suggested and tested, including hydrodynamic conditions, tidal cycle, and high population
density (Palmer 1988).
In this investigation, we test if swimming by Paranais
litoralis is a specialized migration behavior and what
the consequences of migration are for the population
biology of the species. Our goal is to link individual
behavior to population processes, a field of study we
believe is particularly important in spatially divided
populations. The study has been done on a species in a
group (oligochaetes) that, to our knowledge, has not
been previously studied from this point of view, but
that shares a common habitat (intertidal mud flats) and
mode of living (endobenthic burrowers) with many
types of organisms around the world. Through a series
of laboratory experiments and field observations of
P. litoralis, we examine the interaction between migration and the morphology and behavior of individuals.
We address the following main questions: (1) What
proximate factors cause migration by swimming, and
can migrators choose where to settle? (2) Do migrators
differ morphologically, physiologically and behaviorally from non-migrators?
METHODS
Stock cultures. Stock cultures of Paranais litoralis
were started in June 1993 from individuals collected in
Flax Pond, an intertidal mud flat surrounded by a Spartina alterniflora marsh on the north shore of Long
Island, New York, USA. New worms were collected
from the field on several occasions during 1993 and
1994, and added to the stock cultures. Cultures were
kept at 13°C in a flow-through seawater system, in
sediment collected from Flax Pond. Sediment was
frozen and kept at –20°C until use. New sediment was
collected every month to ensure good sediment quality.
Induction of migration. Long-term induction experiment: This experiment was designed to investigate
whether migration by swimming was induced by density or by the depletion of resources. We added 3 ml of
frozen and thawed Flax Pond spring sediment and
5 worms from the stock cultures to each of 6 glass
bowls (diameter 55 mm). The bowls had a rim that
extended about 1 cm above the sediment surface, so
that worms had to swim in order to escape from the
bowls. The bowls were put individually in aquaria with
filtered seawater, and kept at 18°C. Under these culture conditions, worm numbers may double every 4 d
if food is superabundant (Nilsson et al. 1997). There
were 2 treatments: resource renewal and resource
depletion. Three of the bowls received new sediment
at each check and therefore should not have experienced resource depletion (but see ‘Discussion’); the
other 3 bowls received the original sediment so that
resources would be depleted with time. Every 3 d we
counted the number of worms in the bowls and the
number of worms on the bottom of the aquaria. The
worms outside the bowls were called migrants. On 3
Nilsson et al.: Migration of oligochaetes
occasions (Days 10, 17 and 27), 10 to 25 randomly
chosen worms from each treatment (sediment renewal
and depletion) and of each category (migrants and
non-migrants), i.e., a total of about 40 to 100 worms per
occasion, were measured (see below). The experiment
was allowed to run until all worms in the resource
depletion treatment were dead (30 d). To check that
migration was not induced by the buildup of excretion
products in the sediment, we washed the sediment in
the resource depletion treatment in ca 50 ml filtered
seawater each time the worms were checked.
Short-term induction experiment: We also tested if
nutrient-poor sediment would induce migration in
worms in good nutritional status. We took the sediment
from the resource depletion treatment, rinsed it with
filtered seawater several times, put it in 5 bowls (diameter 55 mm), and allowed it to stand in seawater with
aeration for 1 wk. New Flax Pond sediment (frozen and
thawed) was added to 5 other bowls which were
treated identically. After 1 wk, the bowls were transferred to individual aquaria and 10 worms from the
stock cultures were added to each bowl. The number
of migrant and non-migrant worms was then counted
after 2 and 5 d without changing the sediment.
Survival and reproduction of migrators. This experiment was designed to investigate possible differences
in life history characteristics between migrant (swimming) and non-migrant worms. In the long-term induction experiment described above, 2 additional
bowls where sediment was not exchanged were set up
in individual trays. After 18 d, all worms found in
the trays (18 swimming worms found outside the bowls
and 14 burrowing worms found in the bowls)
were transferred individually to bowls with new Flax
Pond sediment (frozen and thawed) and cultured
under conditions as described above. These cultures
were checked every 2 d for 12 d, and survival and the
number of worms producing offspring were recorded.
Selection of sediment by swimming worms. This
experiment was designed to investigate if migrating
(swimming) worms would choose where to settle based
on sediment quality. In each of 5 replicate aquaria, 5
bowls were set up: 2 containing 3 ml of frozen and
thawed sediment from Flax Pond (‘good’ sediment), 2
containing sediment kept from previous experiments
where worm populations had exhausted the resources
(‘exhausted’ sediment), and 1 bowl with a 1:1 mix of
the 2 kinds of sediment. The positions of the bowls
were randomly varied among aquaria. Worms (15 individuals from the stock cultures) were added to the 1:1mix bowl (placed in the center of the aquarium). The
number of worms in each bowl and in the surrounding
aquarium was checked every 2 d for 1 wk.
Selection of sediment by burrowing worms. This
experiment was designed to investigate if worms
91
choose patches of sediment based on sediment quality
on a smaller spatial scale by means of burrowing. In
each of 4 replicate aquaria, frozen and thawed sediment was added to a large petri dish (diameter 14 cm).
Each dish contained 6 sediment patches (patch diameter = 25 mm): 3 where the sediment was replaced
with nutrient-exhausted sediment and 3 where sediment was removed and then returned, to control for
the disturbance of the sediment (‘good’ sediment). The
arrangement of patches was randomly varied among
aquaria. The dishes were left undisturbed for 24 h. At
the start of the experiment, 100 worms from stock cultures were added to the center of each dish. After 48 h,
the number of worms in each patch was counted.
Morphology of migrators and non-migrators.
Migrating and non-migrating worms were transferred
to petri dishes filled with filtered seawater and filmed
at 20× with a video camera attached to a dissecting
microscope. The video images were transferred to a
Macintosh PowerPC 8100/AV and measured for length
and projected area using the public domain program
NIH Image Version 1.55 (Wayne Rasband, US National
Institute of Health). Average width was calculated as
the projected area divided by the length.
We counted the number of segments to investigate if
migrating and non-migrating worms differed in the
number of segments. Fifteen migrating and nonmigrating worms from the long-term induction experiment were transferred individually to vials and fixed in
70% alcohol. The number of segments bearing setae
was checked under a compound microscope at 200×.
Swimming activity and sinking rates of worms. We
placed worms from the migration-induction experiment
(migrators and non-migrators) 100 mm above the bottom of a 0.5 l glass beaker (diameter 87 mm) filled with
Flax Pond water. We recorded the time until the worm
reached the bottom of the beaker and the proportion of
this time spent swimming. We also measured the length
and width of these worms as described above.
Diffusion experiments. We added sediment (frozen
and thawed) to petri dishes (diameter 14 cm) in a 3 mm
layer. The petri dishes were put in aquaria with Flax
Pond water and left for 24 h. The next day, 50 nonswimming worms from the stock cultures were added
to the center and allowed to move for 15 to 30 min. We
then stopped the experiment, and the sediment was
collected with a pipette from 5 concentric rings. The
distance traveled by a worm found in one of these rings
was calculated as the geometric average of the distance from the center of the dish to the inner and outer
boundary of that ring. The 2-dimensional diffusion
constant D (mm2 min–1) for the movement of worms
was calculated as D = MSD/4t, where MSD = mean
square displacement = the squared distance a worm
moved from the point of origin, averaged over all
92
Mar Ecol Prog Ser 207: 89–96, 2000
worms, and t = duration of the experiment (Okubo
1980, Berg 1993). The experiment was repeated on 5
occasions, each time with 2 new petri dishes and new
worms. Diffusion coefficients for swimming worms
were measured by adding 10 migrator worms to 3
replicate petri dishes as described above, and recording their movements as they were swimming.
Field occurrence of migration. Ten sediment samples were taken with a cut-off syringe (inner diameter
25 mm) to a depth of 10 mm at random locations in Flax
Pond during an outgoing tide (water depth 45 to
60 cm), on May 31, 1995. Samples were preserved in
4% formalin solution with Rose Bengal added. We also
took water samples at 12 randomly chosen locations.
Each sample consisted of 12 l of water, sieved through
a 250 µm nylon mesh. The residue collected on the
mesh was preserved in 4% formalin. We counted the
number of worms found in the sediment and water
samples, and the number of segments on each worm
was counted at 200× under a compound microscope.
Statistical analyses. Data were analyzed with 1-way
and factorial ANOVA, and ANCOVA in SYSTAT for
Macintosh, Version 5.2 (SYSTAT Inc., Evanston, Illi-
nois, USA). Homogeneity of variances was tested with
Cochran’s test (Winer et al. 1991). Homogeneity of
slopes in ANCOVA was tested by examining the interactions between the covariate and other factors. When
repeated measures on the same subject were analyzed, the MANOVA option for repeated measures
was used (von Ende 1993). Differences between
migrators and non-migrators in survival and fecundity
were tested with a log-linear test for independence
(G-test, Sokal & Rohlf 1981).
RESULTS
Induction of migration
Our results indicate that exhaustion of resources,
rather than density of Paranais litoralis alone, caused
the emergence of migrating worms. The proportion
of migrating worms was significantly higher in the resource depletion treatment compared to the resource
renewal treatment (Fig. 1, repeated-measures ANOVA
time × sediment interaction, F4,19 = 3.145, p = 0.038), although the density of worms was consistently
higher in the latter. After 10 d, migrators also
started to emerge in bowls where sediment
was renewed. It is possible that 3 d between
changes of sediment was too infrequent to
prevent some resource depletion at the very
high densities encountered near the end
of the experiment, or density in itself may
have caused migration at very high densities.
The short-term induction experiment showed
the same pattern (Fig. 1): worms in the resource depletion treatment migrated significantly more than worms in the sediment renewal treatment (repeated-measures ANOVA
time × sediment interaction, F2, 7 = 11.676,
p = 0.006).
We found no significant differences between swimmers and non-swimmers in number of offspring per worm (migrators: 1.25 ±
0.25 mean ± SE, n = 18; non-migrators: 1.0 ±
0.33, n = 14; 1-way ANOVA, F1, 22 = 0.37,
p = 0.54), proportion of worms reproducing, or
in survival for the 12 d period that we followed
individual worms (Fig. 2).
Selection of sediment
Fig. 1. Paranais litoralis. Number of worms and proportion of migrators
in cultures where sediment was renewed (density limitation) and cultures where sediment was not renewed (resource limitation), in longterm experiment (30 d, top panels) and in short-term experiment (5 d,
bottom panels). Mean ± SE, n = 3
This experiment showed that the worms
settled selectively in ‘good’ sediment after
migration by swimming (Fig. 3). Already after
2 d, the number of worms in ‘good’ (i.e.,
93
Nilsson et al.: Migration of oligochaetes
Table 1. Paranais litoralis. Width of worms. ANCOVA showing effect of worm type (migrator or non-migrator, fixed factor), culture condition (sediment renewal or not, fixed factor),
and worm length (covariate)
Fig. 2. Paranais litoralis. Survival and proportion of worms
reproducing in migrators (n = 18) and non-migrators (n = 14)
during a 12 d period following transfer from exhausted sediment to high-quality sediment. Differences among worm
types are not significant (G-test, G = 0.51, p = 0.47)
renewed) sediment was significantly higher than the
number of worms in ‘exhausted’ sediment (2-way
ANOVA sediment effect, F 1, 4 = 96.4, p = 0.006).
Although some of this difference may have been
caused by reproduction of worms in ‘good’ sediment,
the time frame of the experiment was too short and the
difference between treatments too great to be explained by reproduction alone. Furthermore, the number of worms found in the surrounding aquarium was
much lower than would be expected if they had just
settled in proportion to area. This indicates that worms
swam actively until they found a suitable spot. Worms
also chose favorable sediment when burrowing. Significantly more worms were found in spots with ‘good’
sediment than in spots with ‘exhausted’ sediment
(Fig. 3, 2-way ANOVA sediment effect, F 1, 3 = 17.5,
p = 0.025).
Source of variation
df
MS
F
p
Worm type
Culture condition
Worm × Culture
Length
Residual
001
001
001
001
232
0.0051
0.0045
0.0006
0.0047
0.0008
6.18
5.46
0.70
5.66
0.014
0.020
0.400
0.018
Morphological differences
Worms found outside bowls (migrators) were on
average significantly longer than worms that remained
in bowls (1-way ANOVA, F 1, 233 = 13.7, p < 0.001), and
were thinner for a given length (Table 1), irrespective
of the type of sediment. Migrating worms also had
more segments (36.87 ± 2.20, mean ± SE, n = 15) than
non-migrating worms (26.0 ± 1.49) (1-way ANOVA,
F 1, 28 = 16.7, p = 0.0003). Since the fission zone in
Paranais litoralis normally occurs at Segment 18
(Martinez 1993), the average number of segments in
migrants (37) suggests that migrating worms postpone
fission when resources are in short supply and until
they find suitable sediment after migrating.
Swimming frequency and sinking rate
Migrator worms stayed in the water column significantly longer than non-migratory worms (1-way
ANOVA, F 1,115 = 4.49, p = 0.036). Two factors contributed to this: (1) migrators were on average longer
than non-migrators, and longer worms sank slower
than short worms; (2) migrators were more active
swimmers than non-migrators. When the variation
caused by these 2 factors is partitioned out, there is no
significant difference between worm types (Table 2).
Movements in sediment: diffusion experiments
The diffusion coefficient for worms burrowing in
sediment was 4.22 ± 0.36 mm2 min–1 (mean ± SE, n =
10), while migrators swimming had a diffusion coefficient of 34.0 ± 4.57 (n = 3) mm2 min–1.
Fig. 3. Paranais litoralis. Sediment choice by worms while
swimming and while burrowing. Bars show the average number of worms in bowls with renewed or exhausted sediment
(left panel), or in patches of renewed or exhausted sediment
(right panel) after 48 h. Results are pooled for all trays. Means
± SE, n = 10 (swimming) and n = 12 (burrowing)
Occurrence of migration in the field
Swimming worms were found in the water samples
taken in spring of 1995 (1.67 ± 0.93 per 12 l, n = 12). At
94
Mar Ecol Prog Ser 207: 89–96, 2000
this time, worm densities in the sediment were 6.11 ±
6.02 per 10 cm2 (n = 10). The number of segments per
worm was significantly higher for worms found in
water samples (35.2 ± 0.79 segments, mean ± SE, n =
12) than for worms found in the sediment (24.6 ± 0.95
segments, n = 26) (1-way ANOVA, F 1, 36 = 49.4, p =
0.0001). Out of 26 worms, 2 found in sediment samples
had > 30 segments (both had 34 segments), while all
worms found in the water samples had > 33 segments.
The number of segments of swimmers and non-swimmers in the field were not significantly different from
swimmers and non-swimmers found in the laboratory.
DISCUSSION
This study on Paranais litoralis is an attempt to link
the everyday behavior of individuals (feeding) to a
specialized behavior (migration due to shortage of
resources) and the population biology of the species.
We focus our discussion on 3 points: (1) Are the swimming worms really migrators, i.e., is this a specialized
behavior for moving to new habitats? (2) Is migration a
favorable trait? (3) What are the potential consequences for the population biology of Paranais
litoralis?
A crucial point to our discussion is whether migration
in the water column by Paranais litoralis occurs at all in
the field. We did find worms in the water-column samples from May 1995, and these worms were significantly longer than worms found in the sediment, similar to the situation we found for ‘migrators’ in our
laboratory experiments. On average, swimming worms
found in the field had 35.2 segments, which is not significantly different from laboratory migrators, but significantly longer than for non-swimming worms found
in the field and in laboratory cultures. We have also
found worms in colonization plates placed on the sediment surface (Nilsson et al. unpubl. obs.).
Behavioral and morphological characteristics
of migrators
Paranais litoralis is a deposit-feeder, and moves
through the sediment as it feeds. In this way, it has a
natural dispersing mechanism without any swimming
stage. Is the difference between the movement while
burrowing and the swimming behavior we describe
here just quantitative? Dingle (1996) argues that
migration is specialized behavior especially evolved
for moving to new places. Several characteristics of the
swimming worms indicate that they are specialized for
migration. Both in the field and in the laboratory,
migrator worms were longer and thinner than non-
Table 2. Paranais litoralis. Time spent in water column.
ANCOVA showing effect of worm type (migrator or nonmigrator, fixed factor), length (covariate) and time actively
spent swimming (covariate)
Source of variation
df
MS
F
p
Worm type
Swimming activity
Length
Residual
01
01
01
76
00.976
11.737
01.446
00.336
02.90
34.89
04.30
< 0.093
< 0.001
< 0.041
migratory worms. This may simply be a consequence
of swimming itself: worms have to stretch to swim.
However, the difference in the number of segments
(migrators have about 40% more segments) indicates
that the shape difference is real. Being long and thin
has consequences from a hydrodynamic point of view,
as a long thin rod has more hydrodynamic drag than a
short and wide rod of equal volume, and therefore is
more easily transported passively by water currents.
The effect of body shape on actively swimming animals of the size of P. litoralis is difficult to analyze,
because the Reynolds number (based on length and
swimming speeds measured) for swimming P. litoralis
is about Re = 1, a magnitude at which neither viscous
nor inertial forces predominate. In our still-water
experiments, both swimming activity and large size
were important for staying suspended: when worms
were swimming, long worms stayed suspended significantly longer than small worms. This means that there
are at least 2 components to becoming a swimming
migrator: a morphological difference and a behavioral
difference. Furthermore, swimming morphs forego
asexual reproduction and grow to a larger size instead.
From this point of view, there is a qualitative difference
between dispersal via burrowing and swimming. Our
interpretation is that dispersal via burrowing is a necessary consequence of the feeding mechanism of P.
litoralis, while swimming is a change of behavior with
the purpose of moving the individual to a new place as
a response to an external cue (resource exhaustion).
Is migration a favorable trait?
We cannot say if the occurrence of Paranais litoralis
in the water column in the field is due to active swimming or passive resuspension. Armonies (1994) found
P. litoralis in the water column over an intertidal mud
flat in Germany, and he suggested that this was due to
active emergence from the sediment. Certainly many
worms were actively swimming in our experiments,
especially when the population was beginning to crash
owing to exhaustion of resources. P. litoralis is confined
Nilsson et al.: Migration of oligochaetes
to the oxic top surface layer of sediment, so if current
speeds are high it may be difficult for worms to avoid
being swept away by sediment resuspension. However, the mud flat in Flax Pond, where we collected our
material, is probably a low-energy environment: tracks
in the sediment sometimes remain visible for several
tidal cycles, indicating that sediment resuspension is
not extensive, at least not during spring and summer.
Voluntary and involuntary resuspension could provide
some of the advantages that active swimmers have:
resuspension would deposit worms in areas where
sediment is also deposited, which presumably are
areas where food resources are more likely to be abundant. On the other hand, passive resuspension would
give less control over time of emergence and time of
swimming, and thus increase the risk of being washed
away from favorable habitats. Migration of meiofauna
is often an active process in sheltered habitats such as
mud flats in salt marshes (Palmer 1988), but migration
of sediment-burrowing invertebrates may also be a
mix of local and regional processes (Armonies 1994,
Palmer et al. 1996). The most favorable strategy may
be to combine active behavior with passive drift: the
animal may actively emerge into the water column,
passively float with the current, and then actively swim
down to probe the sediment, as has been described for
both stream and marine invertebrates (André et al.
1993, Commito et al. 1995, Palmer et al. 1996).
We expected migrators to be in better physiological
condition than non-migrators, which would give the
former a reproductive advantage when they were
transferred to high-quality sediment. We found no significant differences, however, between migrators and
non-migrators in age at first reproduction or number of
offspring produced, although the statistical power is
low as our sample size was small (a second set of
experiments failed as the cultures became anoxic). In
nature, however, only migrators are likely to find good
sediment at all. Paranais litoralis is able to completely
exploit the resources of a patch, and therefore they
must leave the exploited patch, making the migrator
trait essential for survival. Worms that stay behind
will have little chance of reproduction unless new
resources are added to the patch. Those new resources
must come soon; if the worms starve for long, they cannot be revived to reproduction by new resources (Levinton & Stewart 1988, Nilsson et al. 1997). The fact that
worms could choose ‘good’ sediment over ‘exhausted’
sediment is therefore crucial to the success of migrators.
Population densities of Paranais litoralis in Flax Pond
peak during early summer (Giere & Pfannkuche 1982,
Cheng et al. 1993, Nilsson et al. 1997), and we did find
worms in water samples in May 1995. Armonies (1994)
similarly found P. litoralis in the water column in early
95
summer. Late summer and early autumn may be a critical period for populations of P. litoralis, because
resources are apparently exhausted or at least the sediment is unsuitable for growth (e.g., Cheng et al 1993).
The chances of finding patches of good sediment
nearby is probably low during this period, and it is
therefore important to be able to migrate long distances. Because swimming is likely to lead to increased
predation, the probability of actually finding suitable
new sediment must be sufficiently high to offset this
increased predation risk in addition to the mortality
from settling in an unsuitable patch. In late summer,
the number of potential competitors, e.g., the polychaetes Nereis succinea and various spionids (Levin et
al. 1987, Zajac 1991) is high, and the number of potential predators (juvenile fishes, crustaceans) is also at its
yearly maximum. Migrators are probably less favored
during spring, when resources are still plentiful and
growth can be supported in local patches.
Consequences for population biology
Several characteristics of the biology of Paranais
litoralis contribute to making migration favorable:
(1) complete exhaustion of resources followed by population crashes; (2) ability to choose new habitats during migration; (3) asexual reproduction decreasing the
risk of extinction during an initial colonization phase.
We therefore argue that P. litoralis is an example of an
organism for which migration is not only favorable, but
also essential for the survival of populations. Are the
rates of movement, as measured in our diffusion experiments, fast enough to make any difference to the colonizing ability of P. litoralis? The asymptotic velocity
of invasion of an exponentially growing population
spreading as a travelling wave into a new habitat is
——
√4rD, where r is the growth rate at low population density, and D is the diffusion coefficient (Holmes et al.
1994). Applying this to P. litoralis would give an eventual population invasion rate of 5 cm d–1 in all directions for burrowers, and 15 cm d–1 for swimmers. From
early March to late June (100 d), the offspring from a
single worm spreading as burrowers could colonize an
area of about 80 m2, while swimming migrators could
theoretically occupy an area of 600 m2. This is calculated from rates of movement measured in still water in
our experiments. The rate of movement (and hence
spread) of swimming worms is probably much higher
in moving water in the field. Thus, the difference in
colonizing ability between worms burrowing and
worms swimming is considerable. Assuming an average water depth in Flax Pond at the time of sampling
in May 1995 of 0.5 m, approximately 1 % of the total
population of worms (calculated from the number of
Mar Ecol Prog Ser 207: 89–96, 2000
96
worms per area unit in sediment samples plus water
samples) would be found in the water, and this proportion would easily suffice to spread the population over
the entire mud flat. Previous studies (Giere & Pfannkuche 1982 and references therein) have shown that
P. litoralis is very hard to find during winter, but
extremely abundant during spring. There are several
possible explanations for this: the worms burrow deep
in the sediment during winter, they enter some sort of
resting stage (e.g., overwintering cocoons), or simply
their densities are so low during winter that they are
unlikely to be found. The population growth rates we
have found previously (Nilsson et al. 1997), and the
rates of movement found in the present study indicate
that while a winter resting stage is still possible, it is
not necessary for explaining the seasonal pattern of
P. litoralis. A few randomly scattered individuals at the
start of the spring will suffice to colonize the entire
mud flat in one season.
André C, Jonsson PR, Lindegarth M (1993) Predation on settling bivalve larvae by benthic suspension feeders: the
role of hydrodynamics and larval behavior. Mar Ecol Prog
Ser 97:183–192
Armonies W (1994) Drifting meio- and macrobenthic invertebrates on tidal flats in Königshafen: a review. Helgol
Meeresunters 48:299–320
Berg HC (1993) Random walks in biology. Princeton University Press, Princeton
Cheng IJ, Chang PC (1999) The relationship between surface
macrofauna and sediment nutrients in a mudflat in
Chuwei mangrove forest, Taiwan. Bull Mar Sci 65:
603–616
Cheng IJ, Levinton JS, McCartney M, Martinez D, Weissburg
MJ (1993) A bioassay approach to seasonal variation in the
nutritional value of sediment. Mar Ecol Prog Ser 94:
275–285
Commito JA, Currier CA, Kane LR, Reinsel KA, Ulm IM
(1995) Dispersal dynamics of the bivalve Gemma gemma
in a patchy environment. Ecol Monogr 65:1–20
Dias PC (1996) Sources and sinks in population ecology.
Trends Ecol Evol 11:326–330
Dingle H (1996) Migration. The biology of life on the move.
Oxford University Press, New York
Giere O (1993) Meiobenthology. The microscopic fauna in
aquatic sediments. Springer-Verlag, Berlin
Giere O, Pfannkuche O (1982) Biology and ecology of marine
Oligochaeta, a review. Oceanogr Mar Biol Annu Rev 20:
173–308
Harrison S, Hastings A (1996) Genetic and evolutionary consequences of metapopulation structure. Trends Ecol Evol
11:180–183
Holmes EE, Lewis MA, Banks JE, Veit RR (1994) Partial differential equations in ecology: spatial interactions and
population dynamics. Ecology 75:17–29
Hughes RN (1989) A functional biology of clonal animals.
Chapman & Hall, London
Lasserre P (1975) Clitellata. In: Giese AC, Pearse JS (eds)
Reproduction of marine invertebrates, Vol III. Annelids
and echiurans. Academic Press, New York, p 215–275
Levin LA, Caswell H, DePatra KD, Creed EL (1987) Demographic consequences of larval development mode:
planktotrophy vs. lecithotrophy in Streblospio benedicti.
Ecology 68:1877–1886
Levinton JS, Stewart S (1988) Effects of sediment organics,
detrital input, and temperature on demography, production, and body size of a deposit feeder. Mar Ecol Prog Ser
49:259–266
Martinez DE (1993) On senescence in asexual metazoans.
PhD thesis, Department of Ecology and Evolution, State
University of New York, Stony Brook, New York
Nilsson P, Kurdziel JP, Levinton JS (1997) Heterogeneous
population growth, parental effects and genotype-environment interactions of a marine oligochaete. Mar Biol
130:181–191
Nunney L, Campbell KA (1993) Assessing minimum viable
population size: demography meets population genetics.
Trends Ecol Evol 8:234–239
Okubo A (1980) Diffusion and ecological problems: mathematical models. Springer-Verlag, Berlin
Palmer MA (1988) Dispersal of marine meiofauna: a review
and conceptual model explaining passive transport and
active emergence with implications for recruitment. Mar
Ecol Prog Ser 48:81–91
Palmer MA, Allan JD, Butman CA (1996) Dispersal as a
regional process affecting the local dynamics of marine
and stream invertebrates. Trends Ecol Evol 11:322–326
Sokal RR, Rohlf FJ (1981) Biometry. The principles and practice of statistics in biological research, 2nd edn. Freeman &
Co, New York
Sugg DW, Chesser RK, Dobson FS, Hoogland JL (1996) Population genetics meets behavioral ecology. Trends Ecol Evol
11:338–342
Von Ende CN (1993) Repeated-measures analysis: growth and
other time-dependent measures. In: Scheiner SM, Gurevitch J (eds) Design and analysis of ecological experiments. Chapman & Hall, New York, p 113–137
Winer BJ, Brown DR, Michels KM (1991) Statistical principles
in experimental design. McGraw-Hill, Inc, New York
Zajac RN (1991) Population ecology of Polydora ligni (Polychaeta: Spionidae). II. Seasonal demography variation and
its potential impact on life history evolution. Mar Ecol Prog
Ser 77:207–220
Editorial responsibility: Otto Kinne (Editor),
Oldendorf/Luhe, Germany
Submitted: February 11, 2000; Accepted: June 27, 2000
Proofs received from author(s): October 26, 2000
Acknowledgements. This study was supported by a National
Science Foundation grant to J.S.L. Contribution number 1077
from the Program in Ecology and Evolution, State University
of New York at Stony Brook.
LITERATURE CITED